The human genome includes at least 62 genes for intermediate filaments, cytoskeletal proteins which were once thought to be specific to eukaryotes (Karabinos, 2004). It is now know that crescentin is a bacterial homolog of intermediate filaments which forms part of the cytoskeleton and determines cell shape in the bacteria C. crescentus (Møller-Jensen, 2005; Pogliano, 2008).

    The gene family of intermediate filaments can be divided into five groups: two groups of keratins, vimentins, neurofilaments, and nuclear laminins. Lamins are thought to be the ancestral members of the intermediate gene family in eukaryotes. Nuclear lamins form a mesh beneath the nuclear membrane (Clarke, 2007). They are known throughout eukaryotic kingdoms with the possible exception of unicellular eukaryotes. Invertebrate lamins form a simpler nuclear complex than those of vertebrates (Melcer, 2007).While all eukaryotes possess laminins, the laminins found in vertebrates possess additional domains and an addition of 42 amino acids in one ancestral domain (Reimer, 1998). All vertebrates possess at least lamin B which is found in most cells and lamin A/C which is absent in cells prior to differentiation. Lamin A/C and one fold of the lamin B protein are unique to vertebrates. The lamin B proteins of urochordates are highly specialized compared to those of other deuterostomes (Clarke, 2007). The duplication of intermediate filament genes occurred after the basal deuterostomes had diverged from vertebrate ancestors. Lamin A and vimentin are absent, as are cadherins which link to intermediate filaments (Morris, 2006).

Invertebrate keratins are only distantly related to those of vertebrates. An intermediate form of keratin (thread keratins) linking vertebrate keratins to those of more primitive chordates was first identified in hagfish. Thread keratins were subsequently identified in lampreys, teleosts, and amphibians (Schaffeld, 2006). Hemichordates possess an intermediate filament structure more similar to that of protostomes than to that of chordates (Zimek, 2002).Tunicates possess intermediate proteins homologus to IF subfamilies of keratins, vimentins, and neurofilaments (Karabinos, 2004).   Keratin is expressed in the epithelia of lancelets, beginning in the larval stage (Karabinos, 2001).  Lancelets do possess multiple intermediate filaments which seem to represent a relatively unspecialized ancestral condition (Reimer, 1998).  The epithelia covering the fin of a lancelet is depicted below.


    Placozoans are among the simplest animals and they possess no nervous or muscle tissue.  Their bodies are composed of only 4 types of cells which use actin, microtubules, and intermediate filaments in their cytoskeleton (Grell, from Harrison, Vol.2). 

     The vertebrate cytoskeleton is composed of actin, microtubules, and intermediate filaments.  Actin and tubulin are polymerized through the hydrolysis of ATP while intermediate filaments require additional proteins to form polymers (van den Ent, 2001).  Most of the 50 genes of the intermediate filament gene family are keratins.  While most intermediate filaments form homodimers, keratins form heterodimers of type I and type II keratin filaments.  In amniotes, keratins are only expressed in epithelia while other intermediate filaments, such as vimentin are expressed in mesenchyme.  Amphibians also express keratins in endothelia (such as the cells lining the frog artery in the following image).

 Keratin-like proteins have been identified in primitive chordates (Schaffeld, 2005).  Keratin is known in fish, including jawless fish, although its expression pattern differs from that of tetrapods in that it is known from the goldfish optic nerve, mosquito fish testis, rainbow trout liver, and lamprey nervous system (Conrad, 1998).  The following image is of the lamprey spinal cord.

     The epidermis and nervous system are derived from the same embryonic tissue, neuroectoderm.  In mammals, both these tissue types express intermediate filaments: keratins in skin and a variety of other proteins in the nervous system.  Although lampreys possess glial cells which are morphologically similar to astrocytes, they lack the intermediate filaments normally expressed in astrocytes (vimentin and GFAP) and express keratins.  Lamprey glial cells express keratins in both the brain and spinal cord that are similar to those of the epidermis (Merrick, 1995).   Thirteen keratins are known in amphibian genomes (ten genes of type I keratins and three genes of type II keratins).



    The primary component of the human epidermis and human hair (pictured in the following images) is a group of proteins named keratin

    There are two subfamilies of keratin proteins: the soft keratins which are found in epithelia and around internal organs and the hard keratins of hair and nails/claws (a human hair follicle is pictured below).  Keratin fibers type I fibers must form heteropolymers with keratin type II fibers (Steinert, 1988). Keratins themselves belong to a larger family of intermediate filaments which all contain a helical rods with non-helical regions at either end. 


Urochordates possess a single representative of the type I and type II keratin groups. Gene duplications increased these numbers to 26 and 27 in the human lineage. In terrestrial animals, these keratin genes are organized into two clusters. Therian mammals share a common set of genes which has undergone only minor modifications since ancestral lineages diverged (such as a loss of a gene in rodents and an inactivation of a gene in chimps and humans) (Zimek, 2006).

Of the intermediate proteins in the human genome, there are at least 29 cytokeratins are known, 19 of which are expressed in epithelia and 10 which are expressed in those cells which produce hair.  The cytokeratins of both tetrapods and fish can be grouped into the same 2 groups.  Most vertebrate epithelial cells produce keratin.  Hagfish have only a few keratin genes while higher vertebrates possess many more, presumably the result of duplications early in the vertebrate lineage (Fuchs, 1981; Markl, 1989).

   Most of the acidic keratin genes are located in two closely linked gene clusters on chromosome 17 while most basic keratin genes are located in a cluster on chromosome 12 (Milisavjevic, 1996).  The type I keratin gene cluster includes 27 genes and 4 pseudogenes with cytoplasmic, hair, and inner root sheath genes forming individual clusters.  At least one of these pseudogenes is a functional gene in rodents.   The type II keratin cluster contains 27 genes and 5 pseudogenes.  Mice possess equivalent clusters with virtually all of the same genes in the same order (Hesse, 2004).




     On chromosome 17q12-21 there are 9 genes and several pseudogenes of hard keratins organized into three clusters: KRTHA6--A5--A2--orphan exon--A8--pseudogene--A1--A4--A3B--A3A.


KRTHAP1 is a pseudogene in humans but it is an active gene in chimps and gorillas.  Its loss of function is the result of one base pair change (OMIM).









KRTH2A mutations cause ichthyosis bullosa.

KRTH6A and KRTH6B are the result of a recent gene duplication event.  Mutations in either one can cause pachyonychia congenita.


KRTHB1 mutations cause moilethrix, a disorder which involves hair (and perhaps nails) that breaks easily.





KRTHB6 mutations cause moilethrix.


KRT1 mutations cause epidermolytic hyperkeratosis.

KRT3 is expressed in the cornea (pictured below).

KRT5 is expressed in the normal mammary epithelial cells along with KRT6, 7, 14, and 17 while tumor cells express mainly KRT 8, 18, 19.  Malignant cells of a breast tumor are depicted below.

KRT6 is often expressed with KRT16.

KRT7 is expressed in epithelia.

KRT8 is expressed in fetal tissue, the placenta (pictured below), and other tissues.  Mutations cause cryptogenic cirrhosis in humans and in mice additional problems such as the loss of exocrine tissue and pancreatic abnormalities have been observed.


KRT10 is expressed with KRT1 in epidermal cells superficial to the basal layer.  Mutations cause epidermolytic hyperkeratosis. 

KRT12 is expressed in the cornea and mutations cause Meesmann corneal dystrophy.

KRT13 mutations cause white sponge nevus with its plaques affecting oral mucosa and other areas.

KRT14 mutations cause epidermolysis bullosa.  Basal epithelial cells (pictured below) express keratins 14 and 5.



KRT16 mutations cause pachyonychia.

KRT17 is only expressed in basal epithelial cells, nail beds, and sebaceous glands.  Mutations cause pachyonychia.

KRT18 mutations cause cirrhosis and chronic hepatitis.


KRT19 is the smallest keratin and is expressed in the periderm (a temporary layer which surrounds developing epidermis) and some carcinomas.


KRN1 is an ultrahigh sulfur protein.  One polymorphism is common in Caucasians.

KRN3 mutations cause Meesmann corneal dystrophy.

KRN4 mutations cause white sponge nevus.


KRN23 may be induced in some pancreatic cancers.


There are 16 keratin associated protein genes which are located within the introns of the TSPEAR gene on chromosome 21q22.3 (Shibuya, 2004).



     Laminins are a family of proteins (with a, b, and g subfamilies) which form heterotrimers in the extracellular matrix and form part of the basement membrane. Intermediate filaments include the lamins which are skeletal elements located on the inner surface of the nuclear envelope (karyoskeleton) in eukaryotes. Lamins would have evolved very early in the history of eukaryotes and may represent the ancestral form of all intermediate filaments (Steinert, 1988).




LAMA2 is a component of the basement membrane of the placenta, striated muscle, and Schwann cells.  Mutations can cause muscular dystrophy.

LAMA3 is a component of hemidesmosomes and mutations can cause epidermolysis bullosa.

LAMA4 is produced beneath epithelia, around muscles, and in the perineurium (the perineurium wraps axon bundles in nerves as pictured below).


LAMA5 is expressed in the placenta, heart, lung, skeletal muscle, kidney, and pancreas.

LAMB2 is produced in the glomerular basement membrane (the glomerulus of a frog is pictured below).


LAMB3 mutations cause epidermolysis bullosa.


LAMC2 mutations can cause epidermolysis bullosa.

LAM3 is expressed in the lung, oviduct, epididymis, vas deferens, and seminiferous tubules.


Peripherin is expressed in the peripheral nervous system.


BFSP1 and BFSP2 are expressed in the lens and mutations can cause cataracts.


Desmin is a muscle specific intermediate filament whose mutations can cause muscular weakenss, desmin-related myopathy, and cardiomyopathy.


Vimentin is an intermediate filament of mesenchymal tissue.  Null mutations in mice cause no observable effects other than the absences of vimentin.


GFAP, glial fibrillary acidic protein, is specific to astroglia.  Mutations cause Alexander disease, the first CNS disorder known to be caused by abnormal astroglia.


Fibrillins are non-collagenous fibres found in many connective tissues in vertebrates; mutations in the fibrillin gene can cause Marfan syndrome in humans.  These proteins are also found in the extracellular matrix of cnidarians (Reber-Muller, 1995).



All intermediate filament genes share a conserved a helical region which results in their coiled shape.  The non-neurofilament members of the gene family share at least 6 of eight introns.  Two of the three neurofilament genes (NEF-L and NEF-M) possess homologous introns and were probably derived from duplications of a single ancestral gene (Levy, 1987).


NEFL (neurofilament protein, light polypeptide) mutations have been implicated in Charcot-Marie-Tooth disease, type 2E.  In mice, mutations caused no observable effect other than decreased regeneration in neurons.




NEFH mutations cause susceptibility to amyotrpohic lateral sclerosis.




Connexins and pannexins function in gap junctions. These genes seem to be absent in echinoderm genomes and thus unique to vertebrates (Burke, 2006).